Amazon web services awslogs代理重复创建僵尸进程

我在ubuntu 16上运行

当我启动awslogs代理时，创建了两个进程

$ sudo systemctl start awslogs
root     25973     1  0 Feb13 ?        00:00:00 /bin/sh /var/awslogs/bin/awslogs-agent-launcher.sh --start --background --pidfile --user awslogs --chuid awslogs &
root     25976 25973  0 Feb13 ?        02:16:01 /var/awslogs/bin/python /var/awslogs/bin/aws logs push --config-file /var/awslogs/etc/awslogs.conf --additional-configs-dir /var/awslogs/etc/config

而且，当我停止awslogs时，只有一个进程被杀死。另一个进程仍然存在

$ sudo systemctl stop awslogs
root     25976 25973  0 Feb13 ?        02:16:01 /var/awslogs/bin/python /var/awslogs/bin/aws logs push --config-file /var/awslogs/etc/awslogs.conf --additional-configs-dir /var/awslogs/etc/config

---

当我再次启动awslogs代理时，创建了两个进程。所以整个过程的计数是3。

我遇到了同样的问题

root@ip-192-168-101-176:~# ps -ef |grep aws
root       500 32151  0 16:17 pts/0    00:00:00 grep --color=auto aws
root     31644     1  0 16:02 ?        00:00:00 /bin/sh /var/awslogs/bin/awslogs-agent-launcher.sh --start --background --pidfile --user awslogs --chuid awslogs &
root     31645 31644  0 16:02 ?        00:00:06 /var/awslogs/bin/python /var/awslogs/bin/aws logs push --config-file /var/awslogs/etc/awslogs.conf --additional-configs-dir /var/awslogs/etc/config

进程id“31644”是僵尸进程

我正在考虑是否在crond上使用低于shell的程序

 kill $(ps -ef |egrep "\/var\/awslogs\/bin\/python \/var\/awslogs\/bin\/aws logs push --config-file \/var\/awslogs\/etc\/awslogs.conf --additional-configs-dir \/var\/awslogs\/etc\/config"|awk '{ ppid=$3; if (ppid==1) print $2}')

如果父进程名为1，则此shell程序将终止awslogs进程

我不承担任何责任